characterization of the wave1 knock-out mouse: implications for

Characterization of the WAVE1 Knock-Out Mouse:Implications for CNS Development

John P. Dahl,1 Jeanne Wang-Dunlop,1 Cathleen Gonzales,2 Mary E. P. Goad,3 Robert J. Mark,2 and Seung P. Kwak1

1Department of Molecular Genetics and 2Neuroscience Discovery Research, Wyeth Research, Princeton, New Jersey 08543, and 3Investigative Pathology,Wyeth Research, Andover, Massachusetts 01810

Developing neurons must respond to a wide range of extracellular signals during the process of brain morphogenesis. One mechanismthrough which immature neurons respond to such signals is by altering cellular actin dynamics. A recently discovered link betweenextracellular signaling events and the actin cytoskeleton is the WASP/WAVE (Wiscott-Aldrich Syndrome protein/WASP-familyverprolin-homologous protein) family of proteins. Through a direct interaction with the Arp2/3 (actin-related protein) complex, thisfamily functions to regulate the actin cytoskeleton by mediating signals from cdc42 as well as other small GTPases. To evaluate the role ofWASP/WAVE proteins in the process of neuronal morphogenesis, we used a retroviral gene trap to generate a line of mice bearing adisruption in the WAVE1 gene. Using a heterologous reporter gene, we found that WAVE1 expression becomes increasingly restricted tothe CNS over the course of development. Homozygous disruption of the WAVE1 gene results in postnatal lethality. In addition, theseanimals have severe limb weakness, a resting tremor, and notable neuroanatomical malformations without overt histopathology ofperipheral organs. We did not detect any alterations in neuronal morphology in vivo or the ability of embryonic neurons to formprocesses in vitro. Our data indicate that WAVE1, although important for the general development of the CNS, is not essential for theformation and extension of neuritic processes.

Key words: WAVE1; knock-out; gene trap; CNS development; neuronal morphogenesis; actin dynamics

IntroductionThe mammalian CNS is a highly organized network of cellularinteractions that are formed by neurons as they translate a widerange of extracellular signals into appropriate morphological re-sponses. The role of the actin cytoskeleton in the development ofneuronal morphology has long been established; however, theexact molecular mechanisms underlying this relationship areonly now beginning to be elucidated.

One established link between extracellular signaling eventsand the process of neuronal morphogenesis is the rho subfamilyof GTPases, which includes rho, rac, and cdc42. However, thisfamily of GTPases is not thought to interact directly with the actincytoskeleton; rather, these proteins have been shown to regulateneuronal morphogenesis by effecting downstream componentsof their signaling cascades (Luo, 2000). The identification of theWASP/WAVE (Wiscott-Aldrich Syndrome protein/WASP-family verprolin-homologous protein) family of proteins, whosefunction is to regulate actin polymerization via a direct interac-tion with the Arp2/3 (actin-related protein) complex, has pro-vided a mechanism through which extracellular morphogenicsignals can be translated via rho GTPase signaling into changes inthe cytoskeletal architecture of neurons (Symons et al., 1996;Miki et al., 1998a,b).

To date, five genes encoding individual WASP/WAVE familymembers have been cloned and characterized (Derry et al., 1994;

Miki et al., 1996, 1998b; Bear et al., 1998; Machesky and Insall, 1998;Suetsugu et al., 1999). Members of this family can be subdivided intotwo groups, WASP [WASP and neuronal-WASP (n-WASP)] andWAVEs [WAVE/SCAR1 (suppressor of cAMP receptor defects),WAVE2, and WAVE3], on the basis of conserved N-terminal do-mains (Takenawa and Miki, 2001). However, all members of theWASP/WAVE family contain a C-terminal WCA (WASP-homology, cofilin-binding, acidic domain) region, which has beenshown to be essential for the regulation of the actin cytoskeleton(Machesky et al., 1999; Banzai et al., 2000; Zalevsky et al., 2001). TheWCA region serves as a scaffold to bind monomeric actin via theWH2 domain, and also associate with the Arp2/3 complex at theacidic domain. Through these interactions, the WASP/WAVE fam-ily imparts on the Arp2/3 complex the ability to nucleate actinbranches (Machesky and Insall, 1998; Marchand et al., 2001).

Recently, Lanier et al. (1999) confirmed that proteins func-tioning as intermediates between cellular signaling cascades andthe actin cytoskeleton play an important role in neuronal mor-phogenesis and CNS organization. In fact, mice lacking Mena, aprotein involved in the regulation of growth cone organization,have specific defects in commissural fiber decussation (Gertler etal., 1996; Lanier et al., 1999). However very little is known aboutthe in vivo functions of other such proteins, especially in the CNS.To assess the role of WAVE family members in the developingand adult CNS, we used a retroviral gene trap to disrupt theWAVE1 gene in mice. This strategy allowed us to insert a reportergene under the control of the WAVE1 promoter while function-ally disrupting expression of the WAVE1 mRNA. To study therole of WAVE1 in neuronal morphogenesis, we ascertained thedevelopmental expression of WAVE1 in heterozygous mice, andsubsequently characterized the homozygous knock-out mice.

Received Dec. 3, 2002; revised Jan. 31, 2003; accepted Feb. 3, 2003.We thank Dr. David Howland for critical evaluation of this manuscript, Yijin She for technical assistance, Jianying

Su for generating the primary neuronal cultures, Brenda Lager for contributions to animal husbandry, and Dr.Youping Huang for assistance with statistical analysis.

Correspondence should be addressed to Dr. Seung P. Kwak, Department of Molecular Genetics, Wyeth Research,CN8000 Princeton, NJ 08543. E-mail: [email protected] © 2003 Society for Neuroscience 0270-6474/03/233343-10$15.00/0

The Journal of Neuroscience, April 15, 2003 • 23(8):3343–3352 • 3343

Materials and MethodsNorthern analysis. Northern analysis was performed using commerciallyavailable human multitissue or brain region blots (Clontech, Palo Alto,CA) according to the instructions of the manufacturer. The probe was a375 nt fragment corresponding to nucleotides 2145–2519 on the WAVE1cDNA (GenBank accession number D87459). The probe was generatedby in vitro transcription using T4 polymerase in the presence of 32P dCTP(Amersham Biosciences, Arlington Heights, IL).

Taqman quantitative reverse transcription (RT) PCR. Tissues were ho-mogenized in Triazol (Invitrogen, San Diego, CA) for RNA extraction.RT reaction using Superscript II kit (Invitrogen) was performed follow-ing the recommendations of the manufacturer, with one modification.Before single-strand DNA synthesis, 5 �g of RNA was digested withRNase-free DNase I for 15 min to get rid of any contaminating genomicDNA. Semiquantitative PCR was performed on Taqman 7700(PerkinElmer Life Sciences, Emeryville, CA) in a multiplexing reaction inwhich amplification of WAVE1 mRNA and cyclophilin mRNA (internalstandard) was performed in the same tube. FAM-labeled WAVE1 probe(5�CACCTCCGGCTCCTCTTCAGAT) and VIC-labeled cyclophilinprobe (5� AAGACTGAGTGGCTGGATGGCAAGCATGTGGTC) wereused at 100 nM each (PerkinElmer Life Sciences) and combined with 100nM of flanking WAVE1 primer pairs (forward: 5�CCCAGCCACTGCTT-TGCA, reverse: 5�GGAGGAGCTGGGTGAAGAA) and cyclophilinprimers (forward: 5� TCCCAGTTTTTTATCTGCACTGC, reverse: 5�GCCTTCTTTCACCTTCCCAAA).

Experiments were performed simultaneously for all tissues, and eachpoint was sampled in triplicate. Levels of WAVE1 mRNA were comparedamong tissues using liver as a reference tissue. Differences in expressionamong various tissues and transgenic lines were calculated using theformula fold change � 2 (��CT) and expressed as a fold increase relativeto the reference tissue.

Generation of WAVE1 knock-out mice. WAVE1 knock-out mice weregenerated by Lexicon Genetics (The Woodlands, TX) from Omnibankclone OST66260 using methods described previously (Zambrowicz et al.,1998). Once germline transmission had been validated, the WAVE1(�/�) animals were mated with an animal with a C57JBL/6 backgroundresulting in a 129Sv/lex � C57JBL/6 hybrid background. The WAVE1knock-out line was maintained on this hybrid background for subse-quent progeny.

Rapid amplification of cDNA ends-RCR. WAVE1 5� untranslated re-gion (UTR) was cloned by using the Invitrogen GeneRacer kit. TotalRNA (5 �g) from WAVE1 knock-out mice (hippocampus/cortex) wasused in the kit according to the instructions of the manufacturer. Briefly,5 �g of total RNA was dephosphorylated with calf intestinal phosphatase,phenol/chloroform extracted, and ethanol precipitated. The dephospho-rylated RNA was decapped with tobacco alkaline phosphatase, phenol/chloroform extracted, and ethanol precipitated. GeneRacer RNA Oligowas ligated to the 5� end of the RNA, and the resulting mRNA wasreverse-transcribed using SuperScript II RT with a LacZ-specific reverseprimer [LacZ Race R1: CTGGCCTTCCTGTAGCCAGCTTTC (24mer)]. PCR amplification was performed in a 50 �l reaction usingGeneRacer 5� primer and a reverse LacZ primer (LacZ Race R2: GGTGCGGGCCTCTTCGCTATTACG). After 30 cycles of PCR, 20 �l of theamplification reaction was analyzed on a 2% agarose– ethidium bromidegel. It was necessary to perform nested PCR to increase the specificity andsensitivity of the rapid amplification of cDNA ends (RACE) products forthe 5� end of the gene; 1.0 �l of the original PCR amplification was usedas a template for nested PCR. GeneRacer 5� nested and reverse-nestedLacZ primers [LacZ Race R3: AAGTTGGGTAACGCCAGGGTTTTCC(25 mer)] were used under the same cycling parameters. Two specificRACE products were present on a 2% gel and cloned into the pGEMT-easy (Promega, Madison, WI) vector for sequencing.

PCR of WAVE1 5�UTR. PCR amplification was performed in a 25 �lreaction using 0.5 �l of cDNA (wild-type cortex), 2.5 �l of forwardprimer (10 �M), 2.5 �l of reverse primer (10 �M), 0.5 �l of dNTP solution(10 �M each), 2.5 �l of 10� PCR buffer containing 15 mM MgCl2, and 0.5�l of Amplitaq (PerkinElmer Life Sciences). The cycling parameters were4 min at 94°C for 1 cycle, 30 sec at 94°C, 30 sec at 60°C, 60 sec at 72°C for

30 cycles, and 7 min at 72°C for 1 cycle. After PCR, 20 �l of the amplifi-cation reaction was analyzed on a 2% agarose– ethidium bromide gel.

Western analysis. WAVE1 (�/�), WAVE1 (�/�) and wild-type mice[postnatal day 18 (P18)] were killed by cervical dislocation. The brainswere removed, the cerebral cortex and hippocampus were isolated, andprotein was extracted by homogenization in 1� RIPA lysis buffer (1%Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS, 0.15 M NaCl, 0.01 M

sodium phosphate, pH 7.2, 2 mM EDTA, and protease inhibitors) andsubsequent centrifugation at 13,000 � g in an Eppendorff (Westbury,NY) microfuge. Protein concentrations were determined using the DCprotein assay kit (Bio-Rad, Hercules, CA) according to the instructions ofthe manufacturer. Fifty micrograms of total protein from each animalwas separated by SDS-PAGE and subsequently transferred to a nitrocel-lulose membrane. The membrane was blocked for 1 hr in PBS-T (0.1%Tween-20 dissolved in PBS) containing 5% dry milk. The membraneswere incubated in a 1:1000 dilution of an anti-WAVE1 polyclonal anti-body, a 1:200 dilution of an anti-WAVE2 polyclonal antibody (SantaCruz Biotechnology), or a 1:200 dilution of an anti-WAVE3 polyclonalantibody (Santa Cruz Biotechnology) overnight at 4°C. The WAVE1polyclonal antibody was generated against amino acids 435– 451 usingstandard methods (Pocono Rabbit Farm, Canadensis, PA). The blotswere washed for 30 min with PBS-T and then incubated with a 1:10,000dilution of the appropriate peroxidase-conjugated secondary antibody.Blots were visualized using the Enhanced Chemiluminescence Plus kit(Amersham Biosciences). For quantitative analysis, each blot wasstripped and reprobed with a 1:2000 dilution of an anti-actin monoclonalantibody (Chemicon, Temecula, CA). The relative intensities of bandswere determined using the Scion (Frederick, MD) Image software pack-age, and the expression levels of the individual WAVE proteins werenormalized to actin.

�-galactosidase assay. Wild-type, WAVE1 (�/�) and WAVE1 (�/�)mice (P20) were killed by cervical dislocation. The brains were removed,and protein was extracted according to the above protocol.�-galactosidase activity was quantitated by measuring the relative opticaldensity at 420 nm in the presence and absence of the substrateo-nitrophenyl �-D-galactopyranoside (Sigma).

In situ hybridization. Adult mice (C57BL/6) weighing 25–35 gm werekilled by cervical dislocation. Fresh-frozen brain sections (15 �m) weremounted on polylysine-coated slides and processed for in situ histo-chemistry as described previously. A probe was generated from a PCRfragment corresponding to 1270 –1440 of mouse WAVE1 cDNA. Ampli-fication of mouse brain cDNA using primers (forward: 5� TCCGTCT-GCCTTGTCCACTTC; reverse: 5� GGAGGAGCTGGGTGAAGAA) re-sulted in a 170 bp fragment, which was subsequently subcloned togenerate antisense riboprobes using T7 phage polymerase. Riboprobestranscribed in the presence of 33P-UTP were labeled to high specificityand used at 1 � 10 6 cpm/slide.

�-galactosidase histochemistry. WAVE1 (�/�) and wild-type mice(P24) were killed by cervical dislocation. Brains were dissected from themice and fixed in a 4% paraformaldehyde solution for 1 hr. Brains werecut into 1 mm sections using a brain block and washed in PBS (Invitro-gen) containing 2 mM MgCl2 for 1 hr at 4°C. The sections were thenwashed in PBS containing 2 mM MgCl2, 0.01% sodium deoxycholate, and0.02% Ipegal for 30 min at 4°C. Finally, sections were incubated in�-galactosidase staining buffer [5 mM K3Fe(CN)6, 5 mM K4Fe(CN)6, 2mM MgCl2, 0.01% sodium deoxycholate, 0.02% Ipegal, 10 mM Tris-HCl,pH 7.2, and 1 mg/ml 5-bromo-4-chloro-3-indolyl �-D-galactoside] for1–3 hr at 37°C. Sections were then stored in a 10% formalin solution(Sigma) at 4°C indefinitely.

For the developmental studies, male WAVE1 (�/�) mice were matedto wild-type females for 24 hr. At 9, 12, 15, and 18 d after mating, preg-nant female mice were killed by cervical dislocation and the embryoswere harvested. Embryos at embryonic day 9 (E9) were stained whole-mount, whereas E12 and E15 embryos were cut in half sagitally beforestaining according to the above procedure. E18 embryos were embeddedin OCT mounting medium and frozen on dry ice. The embryos were thencut into 30 �M sections on a cryostat, mounted on polylysine-coatedslides, and stained using the above procedure.

Growth rate analysis. Weight measurements of 10 wild-type, 17

3344 • J. Neurosci., April 15, 2003 • 23(8):3343–3352 Dahl et al. • Characterization of the WAVE1 Knock-Out Mouse

WAVE1 (�/�), and 9 WAVE1 (�/�) mice were taken daily from P2 toP22. Six time points (P2, P5, P9, P12, P17, and P22) were selected tocompute weight gains in the five corresponding time intervals. ANOVAwas used to determine the variability in body weights within and betweeneach genotype. To analyze these data as rates of growth, a linear mixedmodel was applied to log10-transformed body weights, which allowed usto interpret the weight changes as changes in the growth rates of thesemice. The p values for testing weight changes over the five intervals, aswell as pairwise comparisons of weight changes between genotypes, wereadjusted by the Bonferroni method.

Pathology analysis. Wild-type and WAVE1 (�/�) mice (P20) werekilled and their tissues harvested and fixed in buffered 10% formalinsolution. Tissues were then trimmed, embedded in paraffin, and stainedby hematoxylin– eosin. A board-certified veterinary pathologist evalu-ated all slides, and the observed lesions were graded for presence andseverity.

Primary neuronal culture. WAVE1 (�/�) females impregnated byWAVE1 (�/�) males were killed by cervical dislocation when the em-bryos were approximately E16. Embryos were removed from the uterusand dissected under a microdissection scope. The cortical tissue wasisolated, washed with HBSS and centrifuged at 1500 � g for 1 min. Thecells were incubated in trypsin solution (0.3 mg/ml trypsin and 250 U/mlDNase dissolved in HBSS) for 15 min at room temperature, after whichtrypsin inhibitor solution (trypsin inhibitor 0.5 mg/ml and DNase 250U/ml dissolved in DMEM 10% FBS) was added. The cells were collectedby centrifugation at 1500 � g for 1 min, resuspended in DMEM 10% FBSand triturated until no cell clumps were visible. Cell density was deter-mined by trypan blue staining and cells were plated out in DMEM 10%FBS at a density of 1 � 10 5 cells/well of a six-well plate. Twenty-fourhours after plating, the DMEM 10% FBS was replaced with DMEMcontaining B27 supplement (Invitrogen).

Cellomics analysis. The primary neurons were fixed in a 4% parafor-maldehyde solution for 15 min 5 d after plating. Cells were stained usingthe neurite outgrowth hit kit according to the instructions of the manu-facturer (Cellomics, Pittsburg, PA) and analyzed using a Cellomics ArrayScan II at a magnification of 10�. For each field analyzed, the Cellomicsneurite outgrowth software package was used to calculate the number ofneurons, neurite outgrowth index, the average number of neurites perneuron, and the average neurite length per neuron. Approximately 70fields were analyzed per genotype. For each parameter measured, thedata from all fields were used to generate an average and standard error.Wild-type and WAVE1 knock-out values were compared using Student’st test ( p � 0.05).

Golgi impregnation. Mice (P21–P24) from WAVE1 knock-out or wild-type were killed by cervical dislocation and their brains washed in 0.9%saline. A rapid Golgi stain was performed according to the specificationsof the manufacturer (FD NeuroTechnologies, Germantown, MD).Briefly, whole brains were treated for silver impregnation for 2 weeks,cryoprotected for 48 hr, and sectioned at 120 �m on a cryostat. Aftersectioning and mounting on gelatin-coated slides, sections were devel-oped, clarified, then coverslipped in resinous medium.

ResultsExpression analysis of WAVE1 mRNAWe initially determined the expression of WAVE1 in adult hu-man and mouse tissues to identify organs that may be affected inWAVE1 knock-out mice. Using Northern analysis, we demon-strated that in humans WAVE1 expression is limited to the CNSand that the WAVE1 mRNA is expressed in all of the brain re-gions examined (Fig. 1A,B). We subsequently analyzed WAVE1mRNA expression in the mouse brain using quantitative RT-PCRand determined that the WAVE1 transcript is similarly expressedin the mouse CNS. Our analysis of WAVE1 mRNA expression inmouse brain revealed that the highest levels are found in thehippocampus, cerebral cortex, and striatum (Fig. 1C). Thesefindings suggested that WAVE1 plays an important role through-out the CNS and reaffirmed our interest in determining the roleof WAVE1 in neuronal physiology.

Gene-trap insertionTo generate a line of mice with a targeted disruption of theWAVE1 gene we searched the Omnibank ES cell library (LexiconGenetics) and were able to identify one embryonic stem (ES) cellclone (OST66260) that had a retroviral gene-trap insertion �20nt upstream of the translation initiation site on the WAVE1mRNA (Zambrowicz et al., 1998). The ES cell clone OST66260was then used to generate a line of WAVE1 knock-out mice. Thegene-trap allele was inherited in the predicted Mendelian man-ner, indicating that the WAVE1 disruption does not result inembryonic lethality.

Additional characterization of the mouse WAVE1 gene wasnecessary to ensure that the gene-trap insertion had actually dis-

Figure 1. Tissue and brain region expression of WAVE1. Expression analysis was performedusing human multitissue ( A) and brain region ( B) Northern blots probed with a radiolabeledWAVE1 cDNA probe (see Materials and Methods). A single transcript �3000 nt in length cor-responding to WAVE1 was restricted to the CNS but was widely distributed within the humanbrain. C, WAVE1 mRNA expression in the mouse brain was analyzed using Taqman quantitativeRT-PCR. Data are expressed as the fold change relative to the WAVE1 mRNA levels in thehypothalamus.

Dahl et al. • Characterization of the WAVE1 Knock-Out Mouse J. Neurosci., April 15, 2003 • 23(8):3343–3352 • 3345

rupted this gene. Using 5� RACE PCR to clone the 5� end of themouse WAVE1 cDNA, we identified two distinct WAVE1 tran-scripts resulting from the alternate use of two independent 5�exons. The gene-trap insertion was located in the intron up-stream of exon 2 in the mouse WAVE1 gene (Fig. 2). Becauseboth transcripts use intron A, the gene-trap insertion led to com-plete inactivation of this allele (Fig. 2).

Analysis of WAVE family protein levelsWestern analysis using a WAVE1-specific polyclonal antibodywas performed on samples isolated from the cerebral cortex andhippocampus of P20 wild-type, WAVE1 (�/�), and WAVE1(�/�) mice to confirm if the WAVE1 gene was functionally dis-rupted (Fig. 3A). Quantitative analysis of these experiments in-dicated that mice heterozygous for the gene-trap insertion ex-pressed �50% less WAVE1 protein than wild-type controls,whereas mice homozygous for the insertion did not have anydetectable levels of WAVE1 protein in either brain region. Usingthis same approach, we were able to measure the expression levelsof the two remaining WAVE family members in the brains ofmice from all three genotypes. In quantifying the protein levelsof both WAVE2 and WAVE3 in the cortex and hippocampus ofwild-type, WAVE1 (�/�) and WAVE1 (�/�) mice, we deter-mined that the expression levels of both of these proteins weresimilar in all three genotypes (Fig. 3B,C). These results indicatedthat WAVE2 and WAVE3 levels are unchanged irrespective ofWAVE1 expression.

Use of �-galactosidase as a surrogate marker forWAVE1 expressionIn addition to disrupting WAVE1 expression, the retroviral genetrap inserted a reporter gene, bacterial �-galactosidase, under thecontrol of the mouse WAVE1 promoter (Fig. 2). The knock-inpotentially allows for the measurement of WAVE1 promoter ac-tivity using �-galactosidase as a surrogate marker. To ascertainthe functionality of this marker, a colorimetric �-galactosidaseassay was initially performed on protein samples taken from thebrains of mice representing all three genotypes (Fig. 4A). Miceheterozygous for the gene-trap insertion exhibited increased�-galactosidase activity over background levels, whereas micehomozygous for the insertion had a sevenfold increase in�-galactosidase activity compared with heterozygous littermates.Thus, it appears that the retroviral gene trap disrupted theWAVE1 gene and simultaneously inserted the reporter gene un-der the control of the WAVE1 promoter. The reciprocal relation-ship between WAVE1 expression and �-galactosidase activity

suggested that, in the WAVE1 knock-out mice, �-galactosidasecould function as a surrogate marker for WAVE1 expression.

To further validate the use of �-galactosidase as a surrogatemarker for WAVE1 in anatomical or developmental studies, wecompared WAVE1 and �-galactosidase expression patterns inthe brains of adult mice. Neuroanatomical colocalization ofWAVE1 and �-galactosidase in the adult animals was confirmedby comparing in situ hybridization performed on brains prepared

Figure 2. Disruption of the mouse WAVE1 gene. Structure of the mouse WAVE1 gene withboxes representing exons and the shaded areas representing the 5� and 3� UTR. Introns arelettered A through H and their approximate lengths are given. The structure of the retroviralgene trap used to disrupt the murine WAVE1 gene is shown with the splice acceptor (SA), the�-galactosidase neomycin resistance gene fusion (�-geo), polyadenylation consensus site(pA), the PGK promoter (PGK), the BTK OST sequence (BTK), and the splice donor site (SD). Thesite of insertion of the gene trap into the WAVE1 locus is marked with an arrow.

Figure 3. Analysis of WAVE1, WAVE2, and WAVE3 protein levels. Western analysis wasperformed on 50 �g of protein extracted from the cerebral cortex and hippocampus of wild-type (Wt) mice as well as from mice heterozygous (Het) and homozygous (KO) for the gene-trapinsertion (n � 3). The analysis was performed using an anti-WAVE1 polyclonal primary anti-body ( A), an anti-WAVE2 polyclonal primary antibody ( B), and an anti-WAVE3 polyclonal pri-mary antibody ( C). All blots were normalized by reprobing with an anti-actin primary antibody.Error bars indicate SEM.


from wild-type mice using a WAVE1 antisense riboprobe andbrain sections from WAVE1 (�/�) mice stained for�-galactosidase (Fig. 5A,B). The in situ hybridization data indi-cated that WAVE1 mRNA is enriched in the cerebral cortex, stri-atum, and hippocampus of adult mice (Fig. 5A). The�-galactosidase staining produced similar results, indicating that�-galactosidase expression in WAVE1 (�/�) mice correlatedwith the neuroanatomical distribution of WAVE1 (Fig. 5B).

Developmental expression of WAVE1To gain insight into the role of WAVE1 during development weperformed a histological analysis of E9, E12, E15, and E18 em-bryos (Fig. 5). In E9 embryos, WAVE1 was expressed throughoutthe entire embryo as measured by �-galactosidase activity (Fig.5). From E12 to E15, the expression of WAVE1 became morerestricted, with the highest levels in the brain and spinal cord.Within the cerebral cortex, the highest levels of WAVE1 expres-sion were found in the cortical plate, with very low levels ofWAVE1 expression detected in the proliferative ventricularzones, indicating that WAVE1 is expressed primarily in differen-tiated neurons (Fig. 5). In addition, at both E12 and E15, periph-eral tissues, including the intestines, lung, heart, and limbs, wereshown to express WAVE1 (Fig. 5). However, the expression ofWAVE1 was almost completely restricted to the brain and spinalcord by E18, reflecting the pattern of distribution found in adults,

as described earlier (Fig. 5). Thus, the expression of WAVE1 be-comes increasingly restricted to the CNS as embryos mature.

Phenotypic characterization of WAVE1 knock-out miceTo examine the biological role of WAVE1 during development aswell as in postnatal animals we characterized the phenotype ofmice homozygous for the gene-trap insertion. At E17 both thewild-type and WAVE1 (�/�) embryos were approximately thesame size, and there were no apparent anatomical abnormalitiesobserved in WAVE1 (�/�) animals (Fig. 6A). WAVE1 knock-out pups were distinguishable from wild-type and heterozygouslittermates on the basis of their smaller size shortly after birth. ByP18 the WAVE1 knock-out mice displayed a severely runted phe-notype (Fig. 6B). In addition, WAVE1 knock-out mice also ex-hibited a resting tremor and hindlimb weakness.

The phenotypic defects of WAVE1-deficient mice may stem inpart from CNS dysfunction in that whole-mount analysis ofbrains from wild-type, WAVE1 (�/�), and WAVE1 (�/�) miceat P18 revealed a reduction in overall brain size as well as amarked decrease in the length of the cerebral cortex in theWAVE1 (�/�) mice (Fig. 6C). In wild-type and WAVE1 (�/�)mice the cerebral cortex extended caudally and covered most ofthe tectum, whereas in the WAVE1 (�/�) mice the cortex wasgenerally reduced, leaving a large portion of the midbrain visibleon inspection. The neuroanatomical abnormalities seen in theWAVE1 (�/�) mice most likely are not attributable to a gener-alized failure in brain development, because other brain regions,including the cerebellum, appear to have developed properly(Fig. 6C). It is interesting to note that although the cerebral cortex

Figure 4. �-galactosidase as a surrogate marker for WAVE1. A, �-galactosidase activity inall three genotypes was quantified using o-nitrophenyl �-D-galactopyranoside (ONPG) as acolorimetric substrate. An inverse relationship between WAVE1 protein levels and�-galactosidase activity was observed. B, The CNS expression pattern of WAVE1 in adult micewas determined using in situ hybridization with a radiolabeled riboprobe directed against the3� UTR of the WAVE1 mRNA. WAVE1 mRNA is particularly enriched in the cerebral cortex (Ctx),striatum (Str), and hippocampus (Hipp). C, The CNS expression pattern of �-galactosidase wasgenerated by incubating brain sections from adult mice heterozygous for the gene-trap inser-tion in a �-galactosidase staining buffer. An overlapping expression pattern was observedbetween �-galactosidase and WAVE1 mRNA.

Figure 5. Developmental expression of WAVE1. Embryos heterozygous for the gene-trapinsertion were harvested at E9, E12, E15, and E18 and stained using a �-galactosidase stainingbuffer. Although WAVE1 is expressed throughout the E9 embryo, localization is more restrictedin the older embryos. At E12 and E15, the expression of WAVE1 is highest in the brain and spinalcord. In the cortex of both E12 and E15 embryos WAVE1 expression is enriched in the outercortical layers, including the cortical plate (CP), whereas it almost completely absent from theventricular zone (VZ). In addition, at both of these ages WAVE1 expression is seen in peripheralorgans, including the lung, intestine, heart, and limbs. By E18, WAVE1 expression is restricted tothe CNS.


forms relatively early in CNS develop-ment, the cerebellum is one of the laststructures to form, suggesting that the roleof WAVE1 may be most critical during theearlier stages of CNS development. Alter-natively, WAVE1 may have an impact onthe development of specific brain regions,where its expression is particularly en-riched, including the cerebral cortex,without affecting the development ofother regions, such as the cerebellum, de-spite lower levels of WAVE1 expression inthese tissues.

The morphological defects of theWAVE1 knock-out mice were associatedwith postnatal lethality. We monitoredthe lifespan of a cohort of wild-type,WAVE1 (�/�) and WAVE1 (�/�) miceand observed that all mice homozygousfor the gene-trap insertion died between P21 and P26, with anaverage lifespan of 23.6 d (Fig. 7A). Lethality among wild-type orWAVE1 (�/�) mice was not observed during the postnatalperiod.

To examine the events leading to lethality in WAVE1 knock-out mice further we examined the postnatal weight gain of allthree genotypes. Daily body weight measurements were recordedand subsequently subdivided into five segments (P2–P5, P5– P9,P9 –P12, P12–P17, and P17–P22) for analysis. The average bodyweights at the six representative days are shown in Figure 7B. Inexamining weight data from the earliest time point, we identifiedsignificant differences in the starting weight between wild-typeand WAVE1 (�/�) mice as well as between WAVE1 (�/�) andWAVE1 (�/�) mice. Because the initial weight of WAVE1knock-out mice was already lower, we compared the growth ratesof these mice by applying a linear mixed model to log10-transformed body weights. This analysis demonstrated that for allfive time intervals examined, wild-type and heterozygous miceshowed significant ( p � 0.0001) weight gains. In contrast,WAVE1 knock-out mice gained weight in the first four timeintervals but lost weight during the final interval (Fig. 7B). Inaddition, our growth-rate analysis indicated that there were nosignificant ( p � 0.272) differences in the weight change betweenwild-type and heterozygous mice, signifying that mice with thesetwo genotypes grew at similar rates. WAVE1 knock-out micegrew at a significantly ( p � 0.0001) reduced rate compared withboth wild-type and heterozygous animals. This study demon-strated that the runted phenotype displayed by the WAVE1knock-out mice is attributable to a significantly reduced growthrate throughout the entire life of these animals. It is interesting tonote that during the final interval (P17–P22), the pups becameincreasingly dependent on solid food. During this same period,the WAVE1 knock-out mice displayed a negative rate of weightgain, suggesting that these animals suffered from a reduction innutrient intake. Not surprisingly, the health status of WAVE1knock-out mice deteriorated during this final interval, with all ofthe knock-out mice dying within 4 d of the last weight measure-ment (Fig. 7A).

The peripheral organs of WAVE1 knock-out mice were exam-ined histologically in an attempt to identify pathology that mayhave contributed to the untimely death of these mice. Overall, thetissues and organs of the wild-type and WAVE1 knock-out micewere very similar in macroscopic and microscopic appearances.The primary difference between the two genotypes was in the size

of the individual tissues, which was not surprising, given therunted phenotype displayed by the WAVE1 knock-out mice. Thekidneys and lungs from WAVE1 knock-out mice were deter-mined to be normal (Fig. 8). The WAVE1 knock-out mice dis-played lengthened intestinal villi compared with wild-type con-

Figure 6. Morphological analysis of WAVE1 knock-out mice. A, At E17 WAVE1 knock-out embryos appear normal and do notdisplay gross morphological abnormalities compared with wild-type littermates. B, WAVE1 knock-out mice at P20 display aseverely runted phenotype compared with wild-type littermates. C, Comparison of brains taken from wild-type mice as well asfrom mice both heterozygous (Het) and homozygous (KO) for the gene-trap insertion reveals a dramatic reduction in the size of thecerebral cortex in the mice lacking a functional WAVE1 gene. Brains from heterozygous mice were indistinguishable from those ofwild-type mice.

Figure 7. Lifespan and growth rate of WAVE1 knock-out mice. A, Individual lifespans ofWAVE1 knock-out mice. The WAVE1 knock-out mice survived 21–26 d after birth, with anaverage lifespan of 23.6 d (n � 13). B, Average weights for wild-type (n � 10), WAVE1 (�/�)(n � 17), and WAVE1 (�/�) (n � 9) mice at P2, P5, P9, P12, P17, and P22. The averageweight of the WAVE1 (�/�) mice is significantly reduced compared with the other two geno-types. *p � 0.001, ANOVA. Within-group comparison reveals a decrease in the weight ofWAVE1 (�/�) mice from P17 to P22. †p � 0.001. Error bars indicate SEM.


trols (Fig. 8), which is indicative of reduced food intake by theWAVE1 knock-out mice because roughage normally wearsthe intestinal villi as it moves through the gastrointestinaltract. The skeletal muscle of the WAVE1 knock-out mice hadan apparent increase in the number of nuclei as well as adecrease in the size of individual myocytes, whereas the hepato-cytes of these animals displayed a normal morphology but weredecreased in size (Fig. 8). The hepatocellular and myofiber atro-phy observed in the WAVE1 knock-out mice, without any addi-

tional lesions detected in those tissues, is consistent with the ideathat the knock-out mice suffered from a lack of nutrient intake.Given that WAVE1 expression appears to be restricted to thenervous system in postnatal mice, it is important to note thatthe myofiber atrophy observed in the WAVE1 knock-out mice isnot typical of nerve-related changes in the muscle structure. Inaddition, although tremor and limb weakness were observedin WAVE1 knock-out mice, these phenotypes did not appear tobe associated with motor neuron degeneration (data not shown).The histopathology of myofibers is consistent with the spinalcord analysis in that the patchy, atrophic bundles often seenin rodent models of neurodegeneration were not observed inWAVE1 knock-out mice. Interpretation of the histopathologicalanalysis along with observations from the growth rate analysisoutlined in Figure 7 suggested that the absence of WAVE1 duringdevelopment has caused a defect in these mice that prevents themfrom being able to survive on a solid-food diet. However,WAVE1 knock-out mice appeared to be capable of the physicalprocess of biting and eating solid food because these mice wereobserved to nibble on food pellets in their cages. Whether CNSdefects, such as perturbations in the hypothalamic regulation ofsatiety, underlie the inability of WAVE1 knock-out mice to sur-vive on a solid-food diet remains to be seen.

Because our WAVE1 expression analysis indicated thatWAVE1 is restricted to the CNS in postnatal animals, histologicalstudies were performed on brains taken from P20 wild-type andWAVE1 knock-out mice. This analysis revealed significant neu-roanatomical abnormalities in the mice lacking WAVE1, includ-ing a thinning of the cerebral cortex and reduction of striatum,lateral septum, and corpus callosum (Fig. 9). The neuroanatomi-cal defects observed in the WAVE1 (�/�) mice correlate withareas that express high levels of WAVE1 as seen by in situ hybrid-ization and �-galactosidase histochemistry, with the exception ofthe hippocampus (Figs. 4B,C, 9). Our analysis indicated that al-though the cerebral cortex is significantly reduced in size, thecortical layers all appear to be present. Although the observedreduction in the corpus callosum did not appear to be attribut-able to the premature termination of decussating axons, it couldpotentially result from fewer neurons sending projections intothis structure. The WAVE1 (�/�) mice also have substantiallyenlarged lateral ventricles, which may result from a decrease inthe volume of the structures surrounding the ventricles. For in-stance, projecting fibers passing through the medial striatum ap-peared to be reduced. Terminal deoxynucleotidyl transferase-mediated biotinylated UTP nick end labeling analysis performedon P20 wild-type and WAVE1 knock-out mice did not showincreased number of cells undergoing apoptosis in the brains ofWAVE1 knock-out animals (data not shown), suggesting thatneuronal cell death does not immediately precede morbidity.Taken together, these data suggest that the neuroanatomical ab-normalities observed in the WAVE1 knock-out mice are attrib-utable to a developmental decrease in the number of cells and/orprojections that constitute the cerebral cortex, striatum, and lat-eral septum. Interestingly, this analysis did not reveal any neuro-anatomical abnormalities in the hypothalamus of the WAVE1knock-out animals.

In vitro analysis of WAVE1 (�/�) neuronal morphologyBased on our initial hypothesis that WAVE1 may play an impor-tant role in neuronal morphogenesis, as well as the idea that theneuroanatomical deficits displayed in the WAVE1 (�/�) micewere attributable to the inability of neurons to form the propermorphology, we examined the growth and morphology of pri-

Figure 8. Histopathology of WAVE1 knock-out mice. Hematoxylin-and-eosin-stained sec-tions were generated from the kidney, lung, skeletal muscle, liver, and small intestine of P20wild-type and WAVE1 knock-out mice. These slides were evaluated for the presence and/orseverity of macroscopic and microscopic lesions. The muscle fibers, liver, and intestinal villi ofWAVE1 knock-out mice appeared different from those of wild-type controls.


mary neuronal cultures from the cerebralcortex of E16 wild-type and WAVE1(�/�) mice. In these primary culture ex-periments, cortical neurons were culturedfor 5 d, after which they were analyzed forvarious indices of neurite outgrowth. Us-ing an automated approach, we were ableto analyze the morphology of 1969 wild-type neurons and 1229 WAVE1-deficientneurons quantitatively. In wild-type neu-rons, the neurite outgrowth index, definedas percentage of neurons positive for neu-rite outgrowth, was 63.13%, whereas neu-rons lacking the WAVE1 protein had aneurite outgrowth index of 68.6% (Fig.10). Wild-type neurons positive for neu-rite outgrowth had an average of 2.85 pri-mary neurites per neuron, whereas knock-out neurons had an average of 2.92primary neurites per cell (Fig. 10). Wewere unable to detect any significant dif-ferences in neurite outgrowth index, thenumber of neurites per cell, or the averageneurite length per neuron between thewild-type and WAVE1 knock-out cultures(Fig. 10). These results were unexpectedgiven the proposed role for WAVE1 inregulating cytoskeletal dynamics and indi-cated that WAVE1 is not essential for neu-rite outgrowth in vitro.

In vivo analysis of WAVE1 (�/�) neuronal morphologyWe addressed whether WAVE1 is required for the formation ofaxons and dendrites in vivo by assessing the morphology of neu-rons in the cortex and hippocampus of wild-type and WAVE1knock-out mice using Golgi impregnation. This analysis indi-cated that cortical neurons from layers III and IV of WAVE1knock-out animals exhibit properly polarized apical dendritesand form multiple, highly ordered dendritic branches (Fig.11A,B). Layers within the hippocampus of WAVE1 knock-outmice were indistinguishable from those of wild-type animals,with pyramidal cells projecting axons toward the fimbria (Fig.11A,B). These results demonstrated that mice lacking WAVE1neurons could still generate the proper morphology, confirmingthat WAVE1 is not essential to the process of neuronalmorphogenesis.

DiscussionWAVE1 is a candidate protein for regulating the process of neu-ronal morphogenesis because of its direct connections to rho–GTPase signaling, which has been implicated in many of the pro-cesses leading to the development of mature neurons, and theArp2/3 complex, a key component in actin polymerization (Luo,2000; Mullins, 2000; Higgs, 2001). To provide insight into thephysiological role of WAVE1 in the CNS, we established a line ofWAVE1 knock-out mice. The data generated using the WAVE1(�/�) mice indicate that although WAVE1 expression is wide-spread in the CNS, it appears to be critical for the proper devel-opment of specific neuroanatomical structures and is not essen-tial for neurite outgrowth.

In characterizing the phenotype of the WAVE1 knock-outmice we determined that the absence of WAVE1 during develop-ment leads to postnatal lethality, with mice homozygous for the

gene-trap insertion surviving for an average of 23.6 d after birth.This information, combined with the wide pattern of WAVE1expression during the early stages of mouse development and theovert postnatal phenotype displayed by mice homozygous for thegene-trap insertion, indicates that WAVE1 plays a crucial role indevelopment. That WAVE1 might play such a role in develop-ment is not surprising given that mice lacking functional

Figure 9. CNS histology of WAVE1 knock-out mice. Brain sections from wild-type and WAVE1 knock-out mice were stainedusing cresyl violet (middle). Those sections are compared with sections from wild-type mice that have been labeled with aWAVE1-specific probe by in situ hybridization (ISH) (left). WAVE1 knock-out mice have enlarged lateral ventricles accompanied byreductions in the lateral septum, striatum, and corpus callosum (top insets). An abnormal reduction of the cingulate gyrus in theretrosplenial cortex underlying the superior colliculus was evident (bottom insets). All of these areas appear to express WAVE1 athigh levels, as determined by ISH. LS, Lateral septum; CC, corpus callosum; Str, striatum; Ctx, cerebral cortex; Cg, cingulate gyrus.

Figure 10. Morphological analysis of WAVE1 cortical neurons. Primary neuronal cultureswere generated from E15 wild-type and WAVE1 knock-out embryos. After 5 d in culture, thecells were fixed and stained. The morphology of the neurons was analyzed at a magnification of10� using the Cellomics Array Scan II and the accompanying neurite outgrowth softwarepackage. Average values are shown �/� the SD. Values generated for wild-type and knock-out neurons were compared using Student’s t test ( p � 0.05). For all of the parameters ana-lyzed, no significant differences between the two populations of neurons were detected.


N-WASP are embryonic lethal and do not survive past E12(Snapper et al., 2001). N-WASP knock-out embryos have defectsin the development of multiple organ systems, indicating thatN-WASP has a more extensive role during development thandoes WAVE1.

In comparison, anatomical defects seen in the WAVE1 knock-out mice are restricted primarily to the CNS. The fact that theneuroanatomical malformations seen in the WAVE1 (�/�)mice occur in brain structures that develop between E12 to E15implies that the biological role of WAVE1 is critical to the processof CNS development during this period. Furthermore, the hip-pocampus and cerebellum, two brain regions that express highlevels of WAVE1 but mature later in development, appear todevelop normally in the WAVE1 (�/�) mice. These observa-tions suggest that WAVE1 has an important role during specificstages of CNS development that coincide with a shift in expres-sion toward a CNS-restricted pattern. The absence of WAVE1expression from the subventricular zones of the developing CNSalso implies that WAVE1 does not regulate the proliferation ofneuronal progenitor cells. In fact, WAVE1 appears to be mosthighly expressed in the outer layers of the developing cortex,indicating that the WAVE1 functions in the later stages of neu-ronal differentiation.

Because our analysis of pathology failed to detect anatomicalmalformations in the WAVE1 (�/�) mice outside of the CNSother than those associated with nutrient intake, it is possible thatadditional WASP/WAVE family members may be compensatingfor the lack of WAVE1 in the periphery. Although the two otherWAVE proteins, WAVE2 and WAVE3, are not highly expressed

in the CNS, they are present in peripheral tissues, and may there-fore offer one potential explanation for the lack of peripheralanatomical malformations in the WAVE1 (�/�) mice (Suetsuguet al., 1999). Within the CNS, expression levels of WAVE2 andWAVE3 proteins remain unchanged in the cerebral cortex andhippocampus, suggesting these two WAVE family members donot compensate for the complete loss of WAVE1 by increasingexpression in the CNS of WAVE1 knock-out animals. Interest-ingly, WAVE1 (�/�) mice are normal despite a 50% reductionin WAVE1 protein levels. Whether the low levels of WAVE2 andWAVE3 are sufficient to augment WAVE1 activity functionallyin these mice remains to be seen. Nevertheless, the overt pheno-type of WAVE1 knock-out mice indicates that the other WASP/WAVE family members are not able to compensate fully for theabsence of WAVE1 in the CNS of the knock-out animals andpoints to WAVE1 function that is, at least in part, distinct fromthe other WASP/WAVE family of proteins.

Our analysis of neurite outgrowth did not reveal any signifi-cant differences between cortical neurons isolated from wild-typeand WAVE1 knock-out embryos across a number of morpho-metric parameters. However, these experiments were conductedusing standard primary culture conditions and did not examinethe morphology of neurons challenged with different growthconditions. Exposing the cultures to factors that effect neuriteoutgrowth or altering the substratum on which the neurons aregrown, may potentially result in morphological differences be-tween the two populations of neurons. Semaphorins would beone class of factors that could potentially elicit a differential re-sponse in neurite outgrowth from wild-type and WAVE1-deficient neurons, in part because the role of these proteins inregulating growth cone dynamics has been linked to the activityof rac1 (Jin and Strittmatter, 1997). Culturing neurons fromwild-type and WAVE1 knock-out embryos in the presence ofgrowth inhibitory ligands such as NogoA and myelin associatedglycoprotein (MAG) may also be of interest. The neuronal recep-tor for both NogoA and MAG, Nogo receptor (NgR), has alsobeen linked to the regulation of neuronal growth cones via rho-GTPase signaling (Woolf and Bloechlinger, 2002). Because sig-nals mediated via NgR or semaphorins have not been linked tothe activity of WAVE1, neurons from the WAVE1 knock-outmice may be applied to investigate the role of WAVE1 in regulat-ing cytoskeletal dynamics in response to these proteins.

Although WASP/WAVE family members are known to playimportant roles in regulating cytoskeletal actin dynamics, theyare not the only known intermediates that link signal transduc-tion cascades to changes in the actin cytoskeleton. Members ofthe Ena/VASP (vasodilator-stimulated phosphoprotein) familyof proteins have been localized to actin-rich structures, includingfocal adhesions and lamellipodia (Reinhard et al., 1992; Gertler etal., 1996). The Ena/VASP family of proteins is known to associatewith various signaling molecules, including Abl and Src tyrosinekinases and the cAMP-dependent protein kinase (Butt et al.,1994; Gertler et al., 1996). The (�) isoform of Mena, the mam-malian homolog of the Drosophila Ena protein, has been shownto stimulate the generation of actin-rich structures when ex-pressed in fibroblasts. Interestingly, Mena(�) is expressed at highlevels in the developing CNS (Gertler et al., 1996). Mena knock-out mice have pronounced defects in the corpus callosum andhippocampal commissure. The defects arise from the inability ofneurons within these structures to send projections across themidline, resulting in prematurely terminated axons accumulat-ing at the cis face of the fissure (Lanier et al., 1999). In contrast,our WAVE1 knock-out mice exhibit a surprisingly normal hip-

Figure 11. In vivo analysis of neuronal morphology in WAVE1 knock-out mice. Brain sectionsfrom wild-type and WAVE1 knock-out mice were processed for Golgi impregnation to assessneuronal morphology in the cortex (Ctx) and the CA1 field of the hippocampus (Hipp). Arrowsindicate apical dendrites in the cortex. III, Cortical layer III; IV, cortical layer IV; V, cortical layer V;SO, strata oriens; PC, pyramidal-cell layer; SR, strata radiatum.


pocampus. The corpus callosum was indeed reduced, although itappeared to result from a reduction in the overall number offibers comprising the corpus callosum, with no evidence of accu-mulated predecussating axons. The differences between the neu-roanatomical malformations detected in the Mena and WAVE1knock-out mice suggests that these two proteins have specific butnonoverlapping roles in the developing CNS.

Recent studies have also demonstrated that interfering withthe function of Ena/VASP proteins results in defects in neuronalmigration. More specifically, embryonic neurons infected with aretroviral construct that causes all members of the Ena/VASPfamily to be sequestered in the mitochondria appears to increasethe rate at which these neurons migrate from the ventricular zoneinto the developing cortex. Inhibiting Ena/VASP protein func-tion caused pyramidal cells infected with the viral construct tomigrate to more superficial layers of the cortex, suggesting thatthis family of proteins functions to regulate the positioning ofneurons in the developing brain (Goh et al., 2002). In the WAVE1knock-out mice, the layers of the cerebral cortex all appear toform normally, despite an overall reduction in tissue volume,indicating that the defects seen in these animals are not attribut-able to alterations in the rates of neuronal migration in the devel-oping cerebral cortex. The difference in phenotypes seen in theMena knock-out mice, the WAVE1 knock-out mice, and themice infected with the dominant-negative Ena/VASP retrovirusare consistent with the suspected differences in the localization ofthese proteins within actin-rich structures and further distin-guishes the function of the Ena/VASP homology domain versusthe SCAR homology domain (Nakagawa et al., 2001).

ReferencesBanzai Y, Miki H, Yamaguchi H, Takenawa T (2000) Essential role of neural

Wiskott-Aldrich syndrome protein in neurite extension in PC12 cells andrat hippocampal primary culture cells. J Biol Chem 275:11987–11992.

Bear JE, Rawls JF, Saxe 3rd CL (1998) SCAR, a WASP-related protein, iso-lated as a suppressor of receptor defects in late Dictyostelium develop-ment. J Cell Biol 142:1325–1335.

Butt E, Abel K, Krieger M, Palm D, Hoppe V, Hoppe J, Walter U (1994)cAMP- and cGMP-dependent protein kinase phosphorylation sites of thefocal adhesion vasodilator-stimulated phosphoprotein (VASP) in vitroand in intact human platelets. J Biol Chem 269:14509 –14517.

Derry JM, Ochs HD, Francke U (1994) Isolation of a novel gene mutated inWiskott-Aldrich syndrome. Cell 78:635– 644.

Gertler FB, Niebuhr K, Reinhard M, Wehland J, Soriano P (1996) Mena, arelative of VASP and Drosophila Enabled, is implicated in the control ofmicrofilament dynamics. Cell 87:227–239.

Goh KL, Cai L, Cepko CL, Gertler FB (2002) Ena/VASP proteins regulatecortical neuronal positioning. Curr Biol 12:565–569.

Higgs HN (2001) Actin nucleation: nucleation-promoting factors are not allequal. Curr Biol 11:R1009 –R1012.

Jin Z, Strittmatter SM (1997) Rac1 mediates collapsin-1-induced growthcone collapse. J Neurosci 17:6256 – 6263.

Lanier LM, Gates MA, Witke W, Menzies AS, Wehman AM, Macklis JD,

Kwiatkowski D, Soriano P, Gertler FB (1999) Mena is required for neu-rulation and commissure formation. Neuron 22:313–325.

Luo L (2000) Rho GTPases in neuronal morphogenesis. Nat Rev Neurosci1:173–180.

Machesky LM, Insall RH (1998) Scar1 and the related Wiskott-Aldrich syn-drome protein, WASP, regulate the actin cytoskeleton through the Arp2/3complex. Curr Biol 8:1347–1356.

Machesky LM, Mullins RD, Higgs HN, Kaiser DA, Blanchoin L, May RC, HallME, Pollard TD (1999) Scar, a WASp-related protein, activates nucle-ation of actin filaments by the Arp2/3 complex. Proc Natl Acad Sci USA96:3739 –3744.

Marchand JB, Kaiser DA, Pollard TD, Higgs HN (2001) Interaction ofWASP/Scar proteins with actin and vertebrate Arp2/3 complex. Nat CellBiol 3:76 – 82.

Miki H, Miura K, Takenawa T (1996) N-WASP, a novel actin-depolymerizingprotein, regulates the cortical cytoskeletal rearrangement in a PIP2-dependent manner downstream of tyrosine kinases. EMBO J 15:5326–5335.

Miki H, Sasaki T, Takai Y, Takenawa T (1998a) Induction of filopodiumformation by a WASP-related actin-depolymerizing protein N-WASP.Nature 391:93–96.

Miki H, Suetsugu S, Takenawa T (1998b) WAVE, a novel WASP-familyprotein involved in actin reorganization induced by Rac. EMBO J17:6932– 6941.

Mullins RD (2000) How WASP-family proteins and the Arp2/3 complexconvert intracellular signals into cytoskeletal structures. Curr Opin CellBiol 12:91–96.

Nakagawa H, Miki H, Ito M, Ohashi K, Takenawa T, Miyamoto S (2001)N-WASP, WAVE and Mena play different roles in the organization ofactin cytoskeleton in lamellipodia. J Cell Sci 114:1555–1565.

Reinhard M, Halbrugge M, Scheer U, Wiegand C, Jockusch BM, Walter U(1992) The 46/50 kDa phosphoprotein VASP purified from humanplatelets is a novel protein associated with actin filaments and focal con-tacts. EMBO J 11:2063–2070.

Snapper SB, Takeshima F, Anton I, Liu CH, Thomas SM, Nguyen D, DudleyD, Fraser H, Purich D, Lopez-Ilasaca M, Klein C, Davidson L, Bronson R,Mulligan RC, Southwick F, Geha R, Goldberg MB, Rosen FS, Hartwig JH,Alt FW (2001) N-WASP deficiency reveals distinct pathways for cell sur-face projections and microbial actin-based motility. Nat Cell Biol3:897–904.

Suetsugu S, Miki H, Takenawa T (1999) Identification of two humanWAVE/SCAR homologues as general actin regulatory molecules whichassociate with the Arp2/3 complex. Biochem Biophys Res Commun260:296 –302.

Symons M, Derry JM, Karlak B, Jiang S, Lemahieu V, McCormick F, FranckeU, Abo A (1996) Wiskott-Aldrich syndrome protein, a novel effector forthe GTPase CDC42Hs, is implicated in actin polymerization. Cell84:723–734.

Takenawa T, Miki H (2001) WASP and WAVE family proteins: key mole-cules for rapid rearrangement of cortical actin filaments and cell move-ment. J Cell Sci 114:1801–1809.

Woolf CJ, Bloechlinger S (2002) Neuroscience: it takes more than two toNogo. Science 297:1132–1134.

Zalevsky J, Lempert L, Kranitz H, Mullins RD (2001) Different WASP fam-ily proteins stimulate different Arp2/3 complex-dependent actin-nucleating activities. Curr Biol 11:1903–1913.

Zambrowicz BP, Friedrich GA, Buxton EC, Lilleberg SL, Person C, Sands AT(1998) Disruption and sequence identification of 2,000 genes in mouseembryonic stem cells. Nature 392:608 – 611.


characterization of the wave1 knock-out mouse: implications for

Documents